Many of the investigations in this genetic epidemiology project arise from observations in families at high risk of cancer or in other etiologic studies. One specific mutation of BRCA1, 185delAG, was found first in several Jewish breast-ovarian cancer families. Subsequently, we reported that 0.9% of Ashkenazi Jews carried the mutation. We then conducted a survey among 5318 Jews in the Washington DC area to estimate the risk of breast and ovarian cancer associated with mutations in BRCA1 and BRCA2 outside of high risk families. We found 2.3% of these volunteers had one of 3 mutations: in BRCA1, 185delAG or 5382insC; in BRCA2, 6174delT. By comparing the cumulative risk of breast, ovarian, or prostate cancer in first degree relatives of mutation carriers to the risk of these cancers in relatives of individuals without the 3 mutations, we estimated the risks of developing these three cancers. The estimated risk of developing breast cancer among mutation carriers by age 70 was 56% for breast cancer, 15% for ovarian cancer, and 16% for prostate cancer. Although the risks were significantly elevated, they were substantially lower than risks estimated from high risk families. These findings imply that other factors, either genetic or environmental, are important in the development of cancer among mutation carriers. Dysplastic nevi were first identified in melanoma-prone families. In a multi-center case-control study of melanoma, we found dysplastic nevi to be a central risk factor for melanoma, conferring a 12-fold increased risk among individuals with >10 dysplastic nevi. An increased number of small nevi in the absence of dysplastic nevi was associated with a two-fold increased risk of melanoma; an increased number of both large non dysplastic and small nevi was associated with a four-fold risk. Children with Beckwith-Wiedemann Syndrome (BWS) have a several hundred-fold increased risk of Wilm's tumor, hepatoblastoma and neuroblastoma. Serial ultrasounds identifying kidneys >95% in size is a strong predictor of risk of Wilms' tumor. Although p57 was considered a potential candidate gene, only a small percentage of individuals with BWS actually have germline mutations in p57.